T 1078/23 (Non-dormant state in SCell groups/XIAOMI) 23-05-2025

1. MAIN REQUEST

Claim 1 of the main request comprises the following limiting features (board's outline):

A method comprising:

(a) receiving, by a wireless device, one or more RRC messages comprising configuration parameters indicating: a first SCell group comprising one or more first cells; a second SCell group comprising a plurality of second cells;

(b) activating, to a non-dormant state, the one or more first cells and the plurality of second cells;

(c) in response to receiving a DCI indicating transitioning the first SCell group to a dormant state: transitioning the one or more first cells to the dormant state; and maintaining the plurality of second cells in the non-dormant state.

1.1 Claim 1 - inventive step (Article 56 EPC)

Starting point and distinguishing features

1.1.1 In Reasons 2.3.1 of the decision under appeal, the examining division considered prior-art document D1 "to be the prior art closest to the subject-matter of claim 1". This document disclosed features (a) and (b).

1.1.2 The appellant submitted that document D1 failed to disclose at least feature (c) of claim 1, since it did not mention the "dormant state". The application as a whole clearly concerned the management of the activated ("non-dormant") and dormant state. So, document D1 - in spite of belonging to the same technical field as the present application - was not an appropriate starting point. Instead, "the skilled person would rather choose document D5 as the closest prior art document, since this document deals with the same technical purpose as the present invention". Thus, to avoid hindsight, document D5 - which explicitly mentioned those states - should be used.

1.1.3 The appellant's argumentation is not convincing, for the reasons set out below.

1.1.4 First, the board considers, in agreement with T 787/17 (Reasons 5.1, last paragraph), T 967/97 (Reasons 3.2, last paragraph), T 1112/19, Reasons 2.1.3) or T 449/23, Reasons 1.1.7, and contrary to e.g. T 2057/12 (Reasons 3.2.2), T 2759/17 (Catchword) or the conclusions drawn in the first-instance decision of the Unified Patent Court UPC_CFI_1/2023 of the Central Division Munich (point 8.6), that no specific justification for the choice of a starting point for the assessment of inventive step is necessary if inventive step is to be denied, since the claimed subject-matter must be inventive over any state of the art according to Article 56 EPC and since it is not the task of the skilled person, who is the person qualified to solve the underlying objective technical problem according to the problem-solution approach, to "choose a document as the closest prior art" (see e.g. T 1450/16, Reasons 2.1.4). Selection criteria such as the "intended purpose" of the claimed subject-matter thus constitute merely a matter of efficiency for the deciding body.

The board adds that a selected starting point may indeed turn out to not be suitable for denying inventive step if, for example, the resulting objective technical problem formulated on the basis of that starting point is an unrealistic or artificial one. But this does not mean that a starting point is to be disqualified as unsuitable right from the outset. Nor is it relevant in that context whether other pieces of prior art, such as document D5, are "relatively closer" to the claimed subject-matter, as argued by the appellant.

1.1.5 Second, document D1 belongs to the application's technical field and proposes a unified framework for configuration/activation/deactivation of CCs (Component Carriers) and BWPs (BandWidth Parts) in NR (New Radio) systems. Specifically, this document discloses that "SCell and associated BWP configuration have been done previously via RRC signalling" and that "gNB activates SCells via group BWP DCI for BWP activation" (see page 4/3, antepenultimate paragraph), i.e. features (a) and (b). The fact that D1 does not disclose the "dormant state", i.e. feature (c), does however not preclude it from being an appropriate starting point for the assessment of inventive step. After all, both the present application and the system of D1 relate to CA (Carrier Aggregation) in NR.

Technical effect and objective technical problem

1.1.6 Starting from document D1, the appellant derived the following objective technical problem from distinguishing feature (c): "how to improve the managing scheme of the states of the SCells [in D1]".

1.1.7 In the board's view, this formulation is too broad (since a huge number of possible ways of "improving" the respective scheme can be envisaged) and does not reflect the actual technical contribution of feature (c) as disclosed by the present application itself. It would rather seem that the use of the DCI together with "an enhanced monitoring mechanism on a DCI to wake-up a cell from a dormant state" (see paragraph [00287] of the application as filed) - though not described in claim 1 - would bring about a "fast, dynamic and scalable transitioning mechanism" (see paragraph [00286]) yet providing power savings

vis-à-vis the "activated state". Thus, the introduction of the DCI mechanism in combination with the "dormant state" should result in a favourable trade-off between power consumption and state-transitioning delays. In the light of this disclosure, the board substantially agrees with the objective technical problem formulated by the examining division, namely "how to reduce the activation latency in a scenario with groups as the one disclosed by D1", even though claim 1 falls short of details in respect of the respective monitoring.

"Could-would" considerations

1.1.8 The appellant submitted that the skilled person would not combine the teachings of D1 and D5 because they were remote documents. They in fact dealt with different methods and algorithms, with different purposes and different targets. The concept of "dormant BWP" of document D5 could thus not be directly applied to the "BWP activation" of D1. If the "dormant state" was introduced in D1, it would be impossible to distinguish exactly what SCell "activation state" and SCell "dormant state" were. Thus, the person skilled in the art could not directly derive, on the basis of D1 and without exercising inventive skill, that the "activation" was the transitioning to "non-dormant state". Hence, document D1 in combination with D5 failed to disclose "activating, to a non-dormant state, the one or more first cells and the plurality of second cells" and "transitioning the one or more first cells to the dormant state". Furthermore, the scenario of "groups" of D1 in combination with the "dormant state" of D5 could not yield the "group-based" SCell state transition as well as "maintenance" according to claim 1. Document D5 almost shared the same opinion that the BWP framework should be enhanced to support the operation with "dormant BWP", and signalling details, including the configuration, activation and deactivation of "dormant BWP", could be discussed later, which in turn meant that the mechanism relative to "dormant BWP" would be further considered, rather than re-using the mechanism for a legacy state (i.e. D1). At any rate, the straightforward combination of D1 and D5 would lead to a method different from the claimed one in which the "dormant state" would still be separately managed for each individual SCell rather than on a "SCell group" basis.

1.1.9 The board disagrees. The "dormant state" described in document D5 in fact addresses the very same problem ("fast transitioning") mentioned in paragraph [00286] of the application (see D5, section 2, Proposal 1: "[...] switching out of dormant-BWP is significantly faster compared to switching out of SCell deactivated state"). The board is also convinced that the skilled person would have indeed combined the teachings of D1 and D5 to solve the aforementioned objective technical problem. The penultimate paragraph of section 2 of D5 in fact teaches that

"[t]he dormant BWP will likely have to be configured by RRC, and the DCI based BWP switching mechanism will have to handle switching into and out of the dormant BWP" (emphasis added)."

The straightforward combination of D1 and D5 would result in the addition of a "dormant-BWP" to the existing "BWP activation states" of D1, namely "activated" and "deactivated". The table of point 2.2.3 of D1 containing "the supported combinations of BWP activation state in association with SCell identification" would thus still be used to carry out "Group Activation of Multiple SCell". To achieve the sought-after technical effect of "fast transitioning", the straightforward combination of D1 and D5 would therefore necessarily lead to the transitioning to the "dormant state" at the "SCell group" level (e.g. using the "group BWP DCI" known from D1).

1.2 In conclusion, the main request is not allowable under Article 56 EPC.

2. AUXILIARY REQUESTS I to III

Claim 1 of the present auxiliary requests differs from claim 1 of the main request in the following added features (board's outline):

(d) transitioning a cell to the dormant state comprises transitioning from a non-dormant BWP of the cell to a dormant BWP of the cell as an active BWP [auxiliary requests I and II];

(e) an active BWP of a cell in the non-dormant state is a non-dormant BWP [auxiliary request II];

(f) each cell of the one or more first cells operates in a first frequency band and/or a first frequency band combination,

each cell of the one or more second cells operates in a second frequency band and/or a second frequency band combination [auxiliary request III].

2.1 Claim 1 - inventive step (Article 56 EPC)

2.1.1 As to features (d) and (e), the appellant submitted that, even if the skilled person did combine the teachings of D1 and D5, the straightforward combination of those documents would at most lead to the maintenance of the scheme of D1 with some additions from D5 for individual SCells, i.e. there would still be "group activation and deactivation" but no transitioning to the "dormant state" of SCell groups.

2.1.2 Concerning feature (f), the appellant submitted that the idea of "frequency-specific grouping" achieved an optimised resource management, e.g. deactivation of high-frequency cells for power-saving while maintaining the low-frequency cells for coverage. This resulted in a more sophisticated and technically improved scheme.

2.1.3 These arguments fail to convince the board for the reasons set out below.

As to features (d) and (e), document D1 already discloses SCell activation/deactivation via BWP signalling. More specifically, section 2.2.2, Proposal 1 of D1 states that

"[...] If any BWP is activated in SCell, SCell is

considered to be activated. If all the BWPs are deactivated, SCell is considered to be deactivated".

These definitions, however, need not change with the introduction of a "dormant-BWP". On the other hand, document D5 discloses transitioning from an

active/non-dormant-BWP to a dormant-BWP. Hence, by straightforward analogy, if no BWP is active and any BWP is dormant in the SCell, the SCell should be considered to be "dormant". Thus, the straightforward combination of D1 and D5 (see point 1.1.9 above) would also directly lead to the application of features (d) and (e) to the system of D1.

Feature (f) cannot render the claimed subject-matter inventive either. In the context of the ongoing 5G NR developments discussed in D1 and D5, different SCells would typically operate different carriers, i.e. different frequencies, be it within the same band (intra-band) or in different bands (inter-band). The advantages suggested by the appellant cannot be credibly derived from the claimed feature "as is" since: (i) the feature generally refers to "a first/second frequency band and/or a first/second frequency band combination" (emphasis added) and (ii) there is no specific guidance as to how the frequency grouping should actually be done.

2.2 It follows that none of the present auxiliary requests is allowable under Article 56 EPC.

3. Since there are no allowable claim requests on file, the appeal must be dismissed.